THE ORIGINS AND DEVELOPMENT OFTHE

ANDEAN STATE

EDITED BY

JONATHAN HAAS

School of American Research, Santa Fe, New Mexico

SHELIA POZORSKI

Pan American University, Edinburg, Texas

THOMAS POZORSKI

Pan American Universuy, Edinburg, Texas

The TIght 0/ ,Ire Unt~.u!I'Y of Canthrldge to prmt and Jell

all manner of books wa.r granted by Henry V/lIJ/l /JJ4 Tnt Unl'l~rllf}l ha.s prlnftd and publlSlttd COIJlIIlIJOW/Y since 1584

CAMBRIDGE UNIVERSITY PRESS

CAMBRIDGE

NEW YORK PORT CHESTER MELBOURNE SYDNEY

The work from which this copy was made did not include a formal Copyright notice. This work may be protected by copyrighi la'N. Uses may be allowed with permission from the rightsholder or If the copyright on the work has expired, or if the use is "fair use" or within another exemption. The user of thIS work IS responsible for determining lawful uses.·

The problem

The Middle Horizon period of Andean prehistory (c A.D. 600 to 1000) witnessed the evolution of a complex political orgamzation centered around the city of Wari in the Ayacucho Basin of the central highlands Although there is little doubt that Wari politically dominated this heartland, there IS much dispute over the extent and nature of Wan influence outside this zone This paper is an attempt to clanfy this situation and offer some understanding of the evolutionary processes which account for the apparent occurrence of Wari cultural traits outside the Ayacucho Basin

This , of course, requires some understand 109 of the development of the Wari state 10 Ayacucho, from the later phases of the Early Intermediate Period (c. A.D. 400 to 600) through Middle Horizon Epoch 1a (c. A D. 600 to 700), Ideally, this involves documenting population growth and changes in settlement patterns reflectmg the growth of an urban system 10 Ayacucho, as well as understanding the apparent interaction between Wan and Nazca (Menzel 1964) and between Wari and Tiwanaku (Cook 1979). While Isbell has touched on some aspects of the problem elsewhere 10 this volume, I would like to separate the Issues of Wan mside the Ayacucho Basin versus Wari outside the Ayacucho Basin, and address three major problems.

FIrSt, do manifestations of Wan exist outside the Ayacucho Basin, and If so, where and 10 what contexts do these

91

Chapter 9

From state to empire: the expansion of Wari outside the Ayacucho Basin

Katharina 1. Schreiber

occur? Some researchers argue that Wari traits are limited to the Wan heartland, others see some limited drstribution outside Ayacucho; while still others see Wan traits as pervasive throughout the entire Andean region north of Lake Titicaca. Second, If Wari manifestations exist outside the Ayacucho Basin, what mecharusrn best accounts for these and their distnbunon? Some would argue that the apparent presence of Wan traits outside Ayacucho was the result of Simple reciprocal trade (Shady and RUIZ 1979), while at the other extreme some argue that these traits were spread as the result of the expansion of an all-powerful mihtary empire. Third, given the mechanism which best accounts for these manifestations. what can be said about the evolutionary processes involved in the development of this complex political system?

My specific temporal concern IS the period beginning With Middle HOrIZon Epoch 1b (c. AD. 700). By this time state-level organizations had probably evolved 10 a number of areas 10 the Andean region, mcluding those centered in Meche on the north coast (Moseley 1978a, T. Topic 1982), Tiwanaku 10 the far south highlands (Browman 1982), and Cahuachi in the Nasca drainage on the south coast (Schreiber n.d.). The central coast and the Callej6n de Huaylas may also have been characterized by complex political organization Therefore, the theoretical problem in the Middle Horizon is not to document the development of the 'pristme' state, but rather to understand why manifestations of one particular state came to have an

Katharina J. Schreiber

apparently widespread distribution throughout portions of the Andean region while other states did not. The question to ask seems to be not, 'Was Wari a state?,' but rather, 'What was Wan doing outside the Ayacucho Basin, and why?' (cf. Isbell and Schreiber 1978).

Wari manifestations and related events of the Middle Horizon

Ceramics

Wari styles and iconography, closely related provincial styles, and local copies of these have been found in vanous contexts throughout the central Andes north of the zone of Tiwanaku domination. Their apparent absence In the south has been widely Interpreted as reflecting a border beween Wari and Tiwanaku (Menzel 1964). The presence of pure and derived Wari styles elsewhere IS explained as the result of anyone of many processes Relatively 'pure' Wari styles occur in the highlands from Cuzco to Cajamarca, and on the south and central coast. The Situation in the far north highlands and on the north coast is less clear; few pure Wan style ceramics are found, although some examples which may be denved or copied from Wari styles do seem to occur.

Architecture

A second mamfestation of Wan outside the Ayacucho Basin IS the occurrence of a number of rectangular structures,

•

• Vlracochapampa

•

•

•

•

• Pachacamac •

oWARI

• Plklllaqta

•

• J Incamocco

•

•

Fig 1 Distnbuuon of sites with Wan architecture

92

ranging in size from 25 to 800 m on a Side, and built in a distinctive and unique architectural style. Clear examples of these structures occur in the highlands from Cuzco and Cajamarca, and possibly on the south and central coast as well (Schreiber 1978). These sites (Figure 1) have been interpreted as storage facilities, tambo (way stations). fortresses, and/or administrative centers. The sirnilanty In style and elements of layout of all of these sites suggests that their construction was directly supervised by architectural specialists from Wari. Regardless of their function or the mechanism through which they came to be distnbuted, the existence of these sites in various areas implies some direct input from Wari in those areas Furthermore, the distribution on these sites is probably the best delimiter of the zone of direct Wan 'Influence.'

Road system

The association of these architectural facilities with particular long-distance roads, many later coopted by the Incas as part of their system of 'royal highways,' Implies the existence of a formal system of roads In the Middle Horizon (Lumbreras 1974a, Schreiber 1978) The association of Wari sites with the roads would indicate that Wari may have been responsible for formalizing and maintaining a system of long-distance roads The labor input necessary in the construction and maintenance of these roads suggests that they were not merely local trails.

Exchange

Recognition of long-distance trade IS certainly not new for the Middle Horizon, but it is worth pointing out that some of this trade In exotics seems clearly associated With Wan Wan particrpation III some sort of long-distance trade network I, indicated by the presence of vanous exouc goods at pure Wan sites, mcludrng gold, bronze, Spondylus sp, turquoise, etc. (Schreiber 1978). In some cases, notably that of turquoise, raw materials were apparently brought to Wan itself, where the finished product, were manufactured. Obsidian was also widely traded since precerarruc time" as has been well documented by Burger and Asaro (1977). They point out that obsidian from the major Ouispisisa source reached its Widest distribution during the Middle Horizon and attribute this to some sort of Wan control

Settlement patterns

In much of the Andean region, especially III the highlands, many local settlement patterns seem to undergo major shifts in the Middle Honzon (Brownian 1976; Schreiber 1982) Various explanations of this change include: environmental change, population growth, the introduction of maize cultivation, and even direct interference by Wari itself I do not suggest here that these changes are a direct mamfestation of Wan outside Ayacucho, but rather that any sigruficant cultural developments dunng the Middle Horizon may be relevant to the understanding of the Wan situation

The expansion of Wari outside the Ayacucho Basin

Other political developments

Among other events of the Middle Honzon, some mention should be made of other major political changes dunng this penod. First, we see the fragmentation and apparent relocation of central authonty in the Moche state from the site of Moche to the site of Pampa Grande further north in the Lambayeque Valley (Day 1972b; T. Topic 1982). Second, we see the growth of several new polities including Capmarca, Sican (Shimada 1982), and Pachacarnac (Menzel 1977). And third, the Nazca culture underwent a degree of fragmentatIOns and the loss of Cahuachi as the center of authonty.

Mechanisms

A number of mechanisms have been suggested In the \\terature that purport to explain the spread of Wan traits outsIde Ayacucho. In this section I discuss three possible mechanisms and bnefly evaluate each of these in terms of the Wari manifestations and events of the Middle Honzon dtscussed above.

Religion

Religion has been suggested by some as the mechanism by which Wari traits came to be distributed throughout much of the Andean region (Menzel 1964). That IS, the Wan 'expansion' was an active proselyuzmg effort to spread a particular set of religious beliefs. Certainly the spread of Wan Iconography, including specific mythical beings, as well as the presence of the offenng tradition In Epoch I, could argue in favor of this mterpretation (cf. Cook 1979) However, If rehgion were the only factor involved, one probably would not expect the presence of the large secular architectural facilities and an associated road system. I suggest that If religion were a factor In the spread of Wan traits outside Ayacucho It was more likely a corollary of some other mechanism, rather than a pnmary mechanism in its own right

Commerce

It has been suggested that Wan was one of several urban systems during the Middle Honzon and that extensive trade between these commercial centers accounts for the distribution of Wan traits outside Ayacucho (Shady and Ruiz 1979). The distnbunon of ceramics, specifically when found In these other urban centers, would support this notion. On the other hand, how would one account for the apparent lack of commerce between these centers and Tiwanaku In the south? Again, the presence of major architectural facilities does not entirely Support the notion that Wan extended outward for trade reasons alone If such were the case, one would not expect the major Investment of the labor and resources manifested in these structures More importantly, the fact that all of these tacihnes were apparently built by Just one of the proposed 'commercial centers' argues against this notion,

The existence of a road system would be consistent with extensIve trade between commercial centers; however, the

93

investment of labor and resources necessary for Its construction and rnamtenance again could be expected to be beyond the means of an individual center. LIkewise, the shift In settlement patterns seen In many areas is not entirely consistent with the idea that Wari was one of several co-equal commercial centers, unless a change in subsistence strategies reflects an increased emphasis on particular products to be exchanged In these networks Furthermore, the growth of new pohuca! centers and the eclipse of others outside the Ayacucho BaSIn could be attnbuted to their position and success in the proposed kind of major trading network

As a whole, the participaticn of Wan lD long-distance trade IS consistent with the Idea that it was a major commercial center, though other kinds of data do not completely confirm this as the sole cause of Wan expansion ancI influence outside the Ayacucho Basin Nevertheless, whatever the mechanism that accounts for Wan traits outside Ayacucho. commerce clearly plays a part

Political expanstonlconquest

A third mechanism that has been suggested to account for the distnbution of Wari traits IS that of political domination. In other words, It has been suggested that Wan may be viewed as a political empire which came to dominate large portions of the Andean region (Menzel 1968, Isbell 1978) The distribution of ceramics outside Ayacucho IS consistent With tlus possibility, although ceramics might also be found outside the boundaries of actual political control

ThI) notion 15 consistent with the distribution of architectural facthties , especially If these are Interpreted as pnmanly adrmrustranve facilities. which may have functioned additionally as tambo ; storehouses, and defensive SItes. Certainly the maintenance of an administrative hierarchy would require the construction of mtrusive adrrunistranve facilities, particularly where local levels of adrnimstration were insufficiently complex to fulfill Imperial needs

A system of well-defined and maintained roads is to be expected In this situation in order to facilitate communication WIthin the domain, to facilrtate movement of rrulitary forces through the domain, and to control the movement of people WIthin the domain. The existence of long-distance trade is also consistent with political dorrunation, though It cannot be taken as strong supporting evidence of the general hypothesis.

Changes In settlement patterns and subsistence, if the result of environmental change or major population growth, may SImply represent local responses to the same factors which led to the rise of Wan In the first place However, In the case with which I am most familiar , a strong argument can be made that major changes In the settlement/subsistence system were a direct result of Wan dorrunation (Schreiber 1982)

FInally, the political changes of the MIddle Honzon are not Inconsistent With the notion of Wan dorrunatron, especially the growth of new polities, such as Cajamarca and Sican, around the apparent penphery of the Wan domain. Indeed, the decentralization of pohtical authority in Nasca dunng the

Katharina J. Schreiber

Middle Horizon might be interpreted as the result of the imposition of political control from Wan; and the apparent fact that Nazca collapsed along with Wan at the end of Middle Honzon Epoch 2b is also consistent with this view.

To summanze, religion alone cannot account for the existence of Wan traits outside Ayacucho Although the commercial center/trade network possibihty IS very attractive, two points argue fairly strongly against thu notion. (1) the presence of major architectural facihnes , and especially the Immense cost of establishing such facihties , and (2) the apparent one-sided nature of this commercial network I would suggest, then, that on the basis of our present state of

94

knowledge and mterpretation , the strongest argument can be made In favor of political domination as the mecharusm which best accounts for the presence of Wan mamfestanons outside the Ayacucho Basm The three possible mechamsms are certainly not mutually exclusive, and rehgious and commercial concerns probably played some part in the overall process

CARAHUARAZO VALLEY Lucanas, Ayacucho, Peru

Scale Kilometers

Contour Interval 1 {)() m

o

2

3

4

5

.& Village

• Hamlet

Fig 2 Carahuarazo Valley settlement pattern dunng the Early J mcarnocco Pha,c

e-=----3 e-=----3 E"'"S3
0 2 3 4 5
• Jmcarnocco • Atvpicat Site
C Small Warl . Hamlet
Compound ,
A Village I Wari
I Highway
I Fig 3 Carahuarazo Valley settlement pattern dunng the Late J incamocco Phase

>

The expansion of Wari outside the Ayacucho Basin

A case in the Carahuarazo Valley

One bnef example may serve to emphasize some of these points. The Carahuarazo Valley IS located 10 the south highlands just over the continental divide from the Nazca drainage, about four days' travel (by foot) from Cahuaclu, and ,lbout SIX from Wan. Although located in the modern dc:partment of Ayacucho, the valley lies a considerable distance from the Ayacucho Basin and did not participate 10 any direct \\,IY in the development of the Wari state before Middle Honzon Epoch lb.

When the Wari presence was first felt, there were seven ,mall villages and seventeen seasonally occupied hamlets in the valley (Figure 2). Dunng the penod of Wari domination a number of major changes occurred 10 the valley (Figure 3). (1) The highest Villages, those located at the Juncture of the herding and tuber growing zones, were abandoned and new villages were established at the juncture of the tuber and grain growing zones (2) The site of Jincamocco, a medium-sized Wan enclosure, was built in Epoch lb. During Its occupation It grew to some 27 ha in extent through the addition of more structures around the outside of the onginal enclosure. (This IS the only case with which I am familiar of a Wan enclosure which did not stand isolated from other structures.) (3) At least three more small Wari enclosures were built, two adjacent to grain zones, the third along the road to Wan (4) Two major stepped and paved roads were built out of the valley to the north and south, and a bridge was probably constructed over the nvers. (5) Much of the valley was terraced at this time, and ungatron systems were probably developed concurrently This fact, along with the downward shift In settlement location, suggests the mtroduction of, or the mtensification of, maize cultivanon

The construction of all of these features must have been extremely demanding 10 terms of labor and resources as well as administrative time. The sheer volume of construction suggests that It was not all done at local initiative or expense, accomplished by particrpauon in a commercial network More likely, It was organized and 'financed' by Wari, and was a direct result of political domination and reorganization of the local system by Wan. And what did Wan get In return? The valley was probably a good stopping-off point on the Journey between Wan and Nazca, and may have served as an Important node in the cornrnumcation network It IS possible that Wan was exploiting a small obsidian source in the valley. However, the analysIs done by Burger and Asaro (1977) shows that although Obsidian from this source was predominant on sites within the valley at all time periods. It has not been found at Wan, and ItS dlstnbutlOn dunng the Middle Horizon does not Imply Wan COntrol. Finally, the construction of terraces and the downward shift In settlement locations might suggest that Increased quantities of maize were grown, and may have served as tnbute payments to Wan.

The evolution of empire

If we accept political dorrunanon as the mechanism which

95

best accounts for the distribution of Wari traits outside the Ayacucho Basin, what can we say about the evolution of such an expanding state? For ease of discussion, I divide this evolution into three successive stages

Initial impetus to expansion

After the evolution of state-level adrnirnstranon and, usually, urban society, there are a number of cases 10 history or prehistory 10 which a political organization rapidly and suddenly expanded the boundaries of ItS control In these cases the area that came to be controlled was vastly larger than the inmal heartland. Furthermore, such 'imperial' states may be contrasted with other states that only gradually expand their borders and do so to a much lesser degree.

What causes a state to expand suddenly in this fashion? A number of suggestions have been offered, including: proselynzmg religion (Menzel 1964), control of trade, environmental or population stress leading to a need to control new resource zones (Paulsen 1976, Isbell 1978), and the dictates of certain laws of heredity (Conrad 1981). In the case of Wari, I suspect the answer to this question can only come from an Increased knowledge of the processes inherent 10 the evolution of Wan as a state 10 Ayacucho (cf Isbell 1980). Certainly the Wan expansion had the effect of spreading ItS religion, controlhng at least some long-distance trade, and controlhng new resource zones Whatever the mitral Impetus, Wari suddenly and rapidly expanded outside of Ayacucho beginning In Middle Honzon Epoch lb.

Expansion and consolidation

The second stage 10 the evolution of empires involves two Interrelated processes: expansion and consolidation The direcnons in which an empire expanded, the order 10 which it expanded, and especially the order 10 which areas were consolidated under irnpe nal rule sometimes can be reconstructed more or less directly from the archaeological remains. The process of expansion generally involves some cornbmauon of diplomacy and rrulitansm. As new regions are included III this expansion, the process of consolidanon begins, and Involves the establishment of economic and political ties between the imperial capital and the new ly incorporated

region The actual marnfestation of this process I, the result of the interplay between a number of complex factors Including general relanons between the two groups, location of the new region withm the empire, the adrmmstrative requirements of the empire, and the level of complexity of the local political system, In the case of Wan the processes of expansion and consolidanon are manifested most clearly in presence of major facihties in some areas, as well as the sequence of their construction.

I do not mean to Imply that reconstruction of imperial expansion is a Simple process In fact, political conquest empires are notonously difficult to see archaeologically, even when one has pnor knowledge of their existence. It IS precisely thrs difficulty that suggests that, at least 10 the Andean region,

Katharina 1. Schreiber

we may be dealing with two rather different forms of expanding states. At the risk of oversimplifying a complex situation on the north coast, In the cases of Moche and Chimu, we see examples of expanding states which might be termed to be characterized by intensive forms of control. That is, these states expanded rapidly to the north and south, controlling each successive valley in turn, and Imposed major visible administrative structures In each (see chapter by Mackey, this volume).

On the other hand, Wari, as well as the Inca empire and possibly Tiwanaku, might be said to exhibit more extensive forms of control. Political control seems to be extended over a much larger area by these pohnes; yet within the total area visible mamfestations of imperial control tend to be unequally distributed. That IS, control IS apparent 10 border areas, areas with significant resources, hostile areas, and areas lacking sufficient levels of complexity for Imperial needs Other areas seem to be totally devoid of any overt manifestations of imperial control, quite unlike the much more pervasive mamfestations of the north-coast pohties. In the highland cases, those areas which appear to lack any sort of imperial control may indeed lack such control; they also may represent areas Within the domain that were passed over because they were not needed by or threatening to the empire; or they may have been ruled through diplornanc alliances With local leaders. This latter

* Baran Grande

*CaJamarca 7

•

•

Fig 4 Probable boundanes of the Wari Empire

96

situation is extremely problematic 10 that I can see no immediate means of evaluatmg It archaeologically.

Given that sites With Visible remains of Wari control have

a rather spotty distnbutron throughout the dornam , it seems most feasible to suggest that the overall extent and distribution of sites with Wari architecture and the associated road system gives us the best approximation of the boundaries of the Wari Empire (Figure 4) In other words, I would argue that the Wan pohncal domain, by the end of Epoch 2b, Included all of the highlands from Pikrllaqta, south of Cuzco, to Yanabamba, just south of Cajarnarca , as well as the south and central coast The order of expansion, judging from the ceramics associated With the Wan sites throughout the empire, proceeded first to the southeast toward Nazca and north to Huamachuco or even Cajamarca Only later were major portions of the central highlands north of Ayacucho Incorporated into the empire, and expansion proceeded to the southeast nearly to the Titicaca Basin

Collapse

The third and final stage in the evolution of empires is, of course, their ultimate collapse It IS a cunous aspect of most empires, at least those of the extensive type, that once they cease to expand they do not seem able to maintain themselves. This certainly seems true of Wan. Furthermore, once this point had been reached, the economic and pohtical organization of Wan had changed to such a degree (It was so geared to an expansionist economy, if you Will) that not only did the empire collapse, but the Wan state within the Ayacucho BaSIn also collapsed, and Wan was abandoned. In addition , those polities which were closely tied to Wan, particularly Nazca and to a lesser degree Pachacamac, also collapsed On the other hand, states on Its periphery, such as Tiwanaku, Cajamarca and Sican, continued to exist, and eventually another complex state/empire arose on the north coast

Many reasons have been proposed to account for the collapse of complex ctvihzations In different parts of the world. As In the case of understanding the first stage of empire evolution, the mitral impetus to expansion, the real reasons for collapse may be extremely complex, and difficult to see In the archaeological record. As a first step, however, I would suggest not asking why empires such as Wari collapsed, but rather, why did such pohties stop expanding? In the case of Wari, the answer to the latter question probably lies tn the existence of other complex pohtical organizations to both the north and south of the Wari domain, which it could not or would not subsume. I suggest that Wan Itself regarded these areas as outside ItS domain, and burlt its two largest Iacihties near these border zones: Viracochaparnpa , just short of the Cajamarca pouty to the north, and Pikillaqta , Just short of the Tiwanaku domain to the south The recognition by Wari of these as the limits of ItS dornam and the failure to continue expanding beyond these limits may have preordained the collapse of the Wari empire.